---
title: 'MIDOG 2025: Mitosis Domain Generalization Benchmark'
url: https://www.emergentmind.com/topics/mitosis-domain-generalization-midog-2025-challenge
type: topic
---

# MIDOG 2025: Mitosis Domain Generalization Benchmark

Searching arXiv for recent MIDOG 2025 challenge papers and the main overview paper.
The **Mitosis Domain Generalization (MIDOG) 2025 Challenge** is the third large benchmark in the MIDOG series and frames mitosis analysis as a problem of **domain generalization** under realistic histopathology variability. It was organized as a MICCAI 2025 satellite event and consists of **two tracks**: **Track 1**, detection of mitotic figures as objects in histology images, and **Track 2**, classification of confirmed mitotic figures into **normal** versus **atypical mitotic figures (AMFs)** [2606.07368]. MIDOG 2025 extends earlier MIDOG editions by evaluating models not only across multiple tumor types, species, and scanners, but also across distinct tissue contexts rather than hotspot-only regions, thereby shifting the benchmark toward what the organizers describe as detection “in the wild” [2606.07368]. In the challenge overview, the test dataset comprises **365 cases**, spanning **12 distinct human, canine and feline tumor types**, with **18 teams** in the detection track achieving **F1 scores ranging up to 0.740**, and **21 submissions** in the AMF classification track achieving **balanced accuracy values up to 0.908** [2606.07368].

## 1. Challenge scope and conceptual framing

MIDOG 2025 was designed to test whether mitosis analysis algorithms can operate robustly across **many tumor types, species, scanners, and tissue regions**, not merely across scanner-induced variation or pathologist-selected hotspots [2606.07368]. This distinguishes it from earlier MIDOG editions: **MIDOG 2021** focused on **scanner-induced domain shift**, while **MIDOG 2022** extended to **multiple tumor types and species** but still restricted evaluation to **expert-selected hotspot regions** [2606.07368]. A central premise of MIDOG 2025 is that clinical deployment requires robustness in the broader histological landscape, including random tissue areas and regions rich in hard negatives [2606.07368].

The challenge is motivated by the clinical importance of **mitotic count (MC)** and mitotic characterization in tumor grading and prognostication. Manual assessment is described as being affected by **inter-rater agreement** limitations and **sampling bias**, especially during hotspot selection [2606.07368]. MIDOG 2025 therefore asks not only whether a model can find mitotic figures, but also whether it can distinguish **normal** from **atypical** mitotic figures, since AMFs are described as emerging prognostic markers [2606.07368].

The formal distinction between the two tracks is stable across the challenge literature. Track 1 addresses **mitotic figure detection** under strong domain shift [2509.02593], whereas Track 2 addresses **binary classification** of mitotic figure crops into **normal mitotic figures (NMFs)** and **atypical mitotic figures (AMFs)** [2508.18831], or equivalently **typical** versus **atypical** mitoses [2508.21035]. A plausible implication is that MIDOG 2025 links two historically separate tasks—object-level mitosis detection and subtype-level mitosis interpretation—within a single generalization benchmark.

## 2. Test set design, domains, and annotation protocol

For **Track 1**, the challenge overview states that the test set contains **122 whole slide images (WSIs)**, each from a different patient, covering **12 tumor domains** and multiple scanners [2606.07368]. For each WSI, the organizers selected **three non-overlapping 2 mm² regions of interest (ROIs)**: a **hotspot ROI**, a **random ROI**, and a **challenging ROI** [2606.07368]. One WSI was too small for non-overlapping random selection, yielding a final total of **365 ROIs**: **122 hotspots**, **121 random**, and **122 challenging** [2606.07368]. The hotspot ROI represents the traditional high-proliferation region; the random ROI is sampled uniformly from tissue regions with at least **80% tissue coverage**; the challenging ROI is selected for high densities of imposters and artifacts such as apoptotic or necrotic cells, hyperchromatic nuclei, inflammatory cells, ink markings, out-of-focus artifacts, tissue folds, and delayed fixation areas [2606.07368].

For **Track 2**, the challenge overview states that the dataset is derived from **MIDOG++**, with **11,939 mitotic figures** across **seven tumor domains**, represented as **128×128 px** patches centered on mitotic figures [2606.07368]. Several Track 2 solution papers describe the official training data in compatible terms. One paper reports **10,191 NMFs**, **1,748 AMFs**, **454 labeled images**, and **9 distinct domains** in the official MIDOG 2025 Track 2 dataset [2509.16935]. Another reports the official training dataset as containing **10,191 instances** of NMFs, **1,748 instances** of AMFs, **11,939** total mitotic annotations, and **454** labeled images, with a test set of **120 cases** covering **12 tumor types** from human and veterinary pathology [2508.18831]. A plausible implication is that different participating teams describe closely related but not always identically enumerated domain partitions of the Track 2 material.

The annotation workflow for **Track 1** used a **PHH3-assisted, multi-expert workflow** [2606.07368]. The H&E slide was scanned, the coverslip removed, the H&E stain washed out, and **PHH3 immunohistochemistry** applied to the same section; H&E and PHH3 WSIs were then registered using an affine transform [2606.07368]. In the first annotation phase, a pathologist classified each candidate into five categories combining PHH3 status and morphology; classes **1–3** were treated as **mitotic figures**, classes **4–5** as **hard negatives** [2606.07368]. A second pathologist then reviewed **30 μm** H&E patches in a blinded phase, with a third expert as tie-breaker in cases of disagreement [2606.07368]. The challenge overview reports **Cohen’s $\kappa \approx 0.48$** for Track 1 and **Cohen’s $\kappa \approx 0.68$** for Track 2 [2606.07368].

For **Track 2**, AMF versus NMF labels were assigned on confirmed MF patches, with two pathologists independently labeling each mitosis according to the morphological criteria summarized from Donovan et al. and a third expert resolving disagreements [2606.07368]. AMFs are described as mitoses with abnormal division patterns such as **multipolar spindles, lagging chromosomes, highly irregular chromatin**, or related abnormalities [2606.07368; 2509.10502].

## 3. Evaluation framework and challenge outcomes

For **Track 1**, the primary metric is the **micro-averaged F1 score** [2509.03614; 2606.07368]. One Track 1 paper explicitly states that MIDOG 2025 evaluates detection with a **micro F1-score**, where a detection counts as a true positive if its centroid lies within **$7.5\,\mu m$** of a ground-truth mitosis [2509.03614]. The challenge overview further reports that performance was analyzed across hotspot, random, and challenging ROIs, and that although most models performed reliably in traditional hotspots, there was **significant performance degradation** in challenging ROIs, where **false positive rates tripled** [2606.07368]. It also states that performance varied significantly across the **12 tumor types**, exposing “blind spots” in current architectures for rare or pleomorphic malignancies [2606.07368].

For **Track 2**, the primary metric is **balanced accuracy** [2508.21035; 2508.18831; 2509.16935]. Several participating papers reproduce the binary definition,
$$
BA = \frac{TPR + TNR}{2},
$$
or equivalent formulations in terms of sensitivity and specificity [2508.18831; 2509.16935]. This metric is repeatedly justified as appropriate for the substantial AMF/NMF class imbalance [2508.21035; 2508.18831]. The challenge overview reports that Track 2 submissions achieved **balanced accuracy values up to 0.908** [2606.07368].

The organizers also evaluated the effect of common inference-time strategies. According to the challenge overview, **ensembling** yielded mean increases of **1.5 percentage points in F1 score** for Track 1 and **1.3 percentage points in balanced accuracy** for Track 2, whereas **TTA showed no relevant improvement** [2606.07368]. This is noteworthy because many participating methods nevertheless relied on either cross-validation ensembles or explicit TTA, suggesting that empirical utility was method-dependent even though the aggregate challenge-level analysis favored ensembling over TTA [2606.07368].

A compact summary of the official challenge-level outcomes is given below.

| Component | Reported value | Source |
|---|---:|---|
| Track 1 participating teams | 18 | [2606.07368] |
| Track 1 top F1 | 0.740 | [2606.07368] |
| Track 2 submissions | 21 | [2606.07368] |
| Track 2 top balanced accuracy | 0.908 | [2606.07368] |

## 4. Methodological landscape in Track 1: detection under multi-context domain shift

The Track 1 ecosystem was dominated by **object detection frameworks**, particularly the **YOLO** family, although segmentation-based and two-stage systems were also represented [2606.07368]. The challenge overview states that among analyzed Track 1 submissions, **10/14** used object detection frameworks and **4/14** used segmentation-based approaches [2606.07368]. Supporting participant papers concretize this diversity.

One YOLO-based paper used **YOLOv12-m** as a **single-stage** detector configured to detect two classes, **true mitosis** and **hard negatives**, trained on **MIDOG++**, **CMC**, and **CCMCT**, and reported **F1-score = 0.801**, **precision = 0.808**, and **recall = 0.794** on the preliminary MIDOG 2025 test set [2509.02593]. That method combined **multi-target Macenko stain normalization**, **human/canine-balanced sampling**, **background tiles**, **NMS**, **test-time augmentation**, and **Weighted Boxes Fusion (WBF)** [2509.02593]. Another Track 1 submission, based on **RF-DETR**, reported **F1 = 0.789**, **recall = 0.839**, and **precision = 0.746** on the preliminary test set and highlighted the importance of **domain-balanced training** and **hard negative mining** while finding that several stain normalization strategies lowered F1 [2509.02599].

Two-stage designs were also common. One paper proposed a **two-stage detection-classification framework** using improved **YOLO11x** proposals followed by a **ConvNeXt-Tiny** classifier and reported **F1-score = 0.882**, compared with **0.847** for improved YOLO11x alone and **0.827** for a basic YOLO11x baseline, with precision increasing from **0.762** to **0.839** and recall remaining comparable [2509.02627]. Another paper used **FCOS** for candidate localization and a **ResNet50 + EfficientNetB2** ensemble for candidate refinement, increasing F1 on MIDOG++ validation from **0.803** for the base detector to **0.8432** with classification refinement [2509.02597]. By contrast, a cautionary report on a **Faster R-CNN** plus three-classifier pipeline achieved **F1-score 0.2237**, **Recall 0.9528**, and **Precision 0.1267** on the challenge test set, and concluded that false-positive suppression remained a fundamental difficulty under diverse domains [2509.02630].

Segmentation-centered approaches likewise appeared. A teacher–student **UNet** system formulated detection as pixel-level segmentation, incorporated **contrastive representation learning** and **domain-adversarial training**, and reported **F1 = 0.7660** in Track 1 on the preliminary test set [2509.03614]. A unified pipeline termed **MitoDetect++** used an **EfficientNetV2-L** encoder within a U-Net-like detector and reserved Track 2 for a separate Virchow2-based classifier [2509.02586]. Team Westwood used **nnUNetV2** for candidate screening followed by a random-forest ensemble of CNN classifiers and obtained **F1 = 0.7450** on the preliminary Track 1 test set [2509.02600].

Taken together, these papers show that Track 1 was not merely a detector benchmark in the narrow sense. It was a benchmark for how to combine **multi-domain training data**, **hard negative handling**, **stain-aware augmentation**, **tiling/fusion policies**, and occasionally **second-stage verification** under a three-context ROI design that strongly penalized background sensitivity [2509.02593; 2509.02627; 2509.02599; 2606.07368].

## 5. Methodological landscape in Track 2: atypical mitosis classification

Track 2 attracted a particularly broad range of classifiers, from conventional CNNs to pathology-specific foundation models. The challenge overview states that most teams fine-tuned or adapted **pre-trained models**, including **LoRA** on vision foundation models such as **UNI**, **Virchow**, **Virchow2**, **HIBOU**, and **DINOv3**, as well as conventional fine-tuning of **ConvNeXt**, **EfficientNet**, **ResNet**, **DenseNet**, and **Hover-Net** [2606.07368].

A recurring design theme was **object-centric restriction of context**. A ConvNeXt V2 submission used **60% center cropping**, **5-fold cross-validation**, and a **ConvNeXt V2 base model**, and reported **balanced accuracy = 0.8831** overall on the development phase, with an ablation showing that center cropping improved balanced accuracy from **0.8090 ± 0.0361** to **0.8492 ± 0.0065** and prevented non-convergent folds [2508.18831]. Another Track 2 paper using **EfficientViT-L2** employed **leave-one-cancer-type-out cross-validation** with **5-fold ensembles**, reporting **balanced accuracy of 0.859**, **ROC AUC of 0.942**, and **raw accuracy of 0.85** in the preliminary evaluation phase [2509.02589].

A second major theme was **multi-task or hardness-aware supervision**. One submission introduced a **multi-task neural network** with a **Pre-Activated ResNet-50** backbone, a classification head, a **mitosis binary segmentation** decoder, and a **pixel-level mitosis classification** decoder, trained with
$$
\mathcal{L} = \mathcal{L}_{cls} + \mathcal{L}_{seg} + \mathcal{L}_{pixcls},
$$
and achieved **balanced accuracy = 0.856** on the preliminary MIDOG25 Track 2 test set [2508.21035]. Another used a **ResNet-50** backbone with **expert-specific phenotype heads** and a **hardness head** to jointly model phenotype and annotation difficulty, reaching **balanced accuracy \(0.8744 \pm 0.0093\)** in 5-fold cross-validation and **overall preliminary BA \(0.8736 \pm 0.0204\)** [2509.10502].

Foundation-model approaches formed a third major cluster. A study of **Virchow**, **Virchow2**, and **UNI** with **LoRA** found that **Virchow with LoRA rank 8** and a 3-fold ensemble achieved **balanced accuracy of 88.37%** on the preliminary test set and **joint 9th place** on the leaderboard [2509.16935]. Another submission based on **UNI2-h** reported that **visual prompt tuning (VPT)** substantially outperformed LoRA, and that adding **Vahadane and Macenko stain normalization** within TTA produced **balanced accuracy = 0.8837** and **ROC-AUC = 0.9513**, ranking within the top 10 teams [2509.02640]. A further foundation-model paper selected **H-optimus-0** with **LoRA**, **soft labels based on multi-expert consensus**, **adaptive focal loss**, **metric learning**, **hard negative mining**, and **domain adaptation**, achieving **mean BA \(0.851 \pm 0.037\)** under leave-one-domain-out evaluation [2509.02601].

More conventional but highly effective CNNs also remained competitive. A **ConvNeXt Small** model trained on all available datasets with a histopathology-specific augmentation pipeline achieved **balanced accuracy = 0.8961** and **ROC AUC = 0.9561** on the preliminary leaderboard [2509.02595]. A **DenseNet-121** framework with **Macenko stain-aware augmentation**, **weighted sampling**, and a **hybrid objective combining class-weighted BCE and focal loss** reported **balanced accuracy 85.0%**, **AUROC 0.927**, **sensitivity 89.2%**, and **specificity 80.9%** on the official test set [2510.22630]. A data-centric UCL submission, grounded in newly released datasets and modern training recipes, reported **Track-2 balanced accuracy of 0.9107** for atypical mitotic cell classification on its own test set [2509.02585]. This suggests that the challenge-level best value of **0.908** [2606.07368] coexisted with participant-side internal evaluations that occasionally exceeded that level on non-identical test partitions.

## 6. Interpretive significance and recurrent lessons

MIDOG 2025 formalized a transition from scanner-centric domain shift benchmarks to **multi-context, multi-tumor, multi-species robustness testing**. The defining insight of the challenge overview is that models that appear reliable in hotspots can degrade substantially in challenging ROIs, with **false positive rates tripling** [2606.07368]. This suggests that the benchmark did not merely test morphology classification in isolation; it tested whether models could resist background-driven or context-driven failure modes in non-curated tissue regions.

Across both tracks, a consistent methodological pattern emerges. High-performing systems generally combined several of the following: **broad domain exposure**, **histopathology-specific augmentation**, **object-centric cropping or localization**, **class-imbalance-aware objectives**, and **ensembling** [2508.18831; 2508.21035; 2509.02593; 2509.02595; 2606.07368]. By contrast, the aggregate challenge analysis found **no relevant improvement** from TTA [2606.07368], even though multiple strong participant systems used it [2509.02593; 2509.02586; 2509.02640]. This is best read not as a contradiction but as evidence that TTA was not a universally reliable lever once averaged over heterogeneous methods and domains.

Another recurring lesson concerns the tension between **foundation models** and **specialized task structure**. Several Track 2 papers show strong results with pathology-specific foundation models adapted by **LoRA** or **VPT** [2509.16935; 2509.02601; 2509.02640], yet other papers report very competitive or stronger performance from carefully tuned CNNs such as **ConvNeXt**, **DenseNet-121**, or **ResNet-50** [2509.02595; 2509.10502; 2510.22630]. This suggests that in MIDOG 2025, large pretrained backbones were advantageous but not sufficient; architectural choice interacted strongly with sampling policy, augmentation strength, thresholding, and the way domain shift was operationalized.

A common misconception is that domain generalization in mitosis analysis is principally a problem of stain normalization. The participant literature gives a more qualified picture. Some methods benefited from **Macenko** or **Vahadane**-based stain augmentation or normalization [2509.02593; 2509.02640; 2510.22630], whereas the RF-DETR paper explicitly reports that **Macenko**, **Multi-Macenko**, **Reinhard**, hematoxylin-only, eosin-only, **CutMix**, and **Gaussian blur** all **decreased F1** in that setting [2509.02599]. This suggests that stain handling is task- and architecture-dependent and that preserving native histomorphological variability can be as important as harmonizing color statistics.

In encyclopedia terms, MIDOG 2025 marks a methodological inflection point. It reframed mitosis analysis from a hotspot-optimized detection problem into a benchmark of **clinical reliability under contextual diversity**, while simultaneously elevating **atypical mitosis classification** into a first-class challenge task [2606.07368]. The resulting body of work shows that the field can now assemble strong multi-domain systems, but also that “in the wild” mitosis analysis remains limited by rare morphologies, hard negatives, and cross-domain blind spots [2606.07368].

Source: https://www.emergentmind.com/topics/mitosis-domain-generalization-midog-2025-challenge